We suppose that the initial events in carcinogenesis result from the induction, as a result of DNA damage, of the synthesis of plasminogen activator which, in turn, results in the activation of "error-prone" processes. We further suppose that the level of plasminogen activator is determined by interaction with the systems producing poly(ADP-ribose) so that failure to produce poly(ADP-ribose)(the natural signal for a shut-off of DNA synthesis in the presence of damage) leads to the formation of plasminogen activator. This hypothesis predicts that protease inhibitors such as antipain will block the error-prone DNA synthesis system and it also predicts an antagonistic interaction between inhibitors of poly(ADP-ribose) synthesis and protease inhibitors since poly(ADP-ribose) synthetase inhibitors should promote error-prone synthesis; whereas, protease inhibitors should block this process. We propose to test this model by biochemical determinations of plasminogen activator and of poly(ADP-ribose) polymerase and of cytotoxicity in human cells, by studies on different modes of DNA synthesis in such cells, and by the development of an in vivo/in vitro model for the study of "error prone" synthesis in which the activity of cell extracts can be assayed and the activity of the inhibitors tested directly.